Knowledge translation (KT) is a growing movement in clinical and health services research, aimed to help make research more relevant and to move research into practice and policy. This paper examines the conventional model of policy change presented in KT and assesses its applicability for increasing the impact of urban health research on urban health policy. In general, KT conceptualizes research utilization in terms of the technical implementation of scientific findings, on the part of individual decision-makers who can be "targeted" for a KT intervention, in a context that is absent of political interests. However, complex urban health problems and interventions infrequently resemble this single decision, single decision-maker model posited by KT. In order to clarify the conditions under which urban health research is more likely or not to have an influence on public policy development, we propose to supplement the conventional model with three concepts drawn from the social science: policy stages, policy networks, and a discourse analysis approach for theorizing power in policy-making.